Ethical approvals

All animal experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Research Council and the Weatherall report, “The use of non-human primates in research”. Animals were housed and experiments performed at the Washington National Primate Research Center (WaNPRC). The University of Washington Institutional Animal Care Use Committee (IACUC) approved the protocol (Permit Numbers: 4165-02, 4165-03, 3328-05). All surgery was performed under general anesthesia and all efforts were made to minimize suffering. Review of human autopsy slides was deemed by the Seattle Children’s Institutional Review Board to not meet the federal definition of “human subjects research.”

Study design, animal groups

Analysis of fetal brain pathology included both pregnant nonhuman primate (NHP) and human samples from 5 cohorts: (1) fetal NHP controls (N = 9), (2) fetal NHP – maternal ZIKV infection (N = 17), (3) fetal NHP – maternal FLUAV infection (N = 10), (4) neonatal NHP subjected to perinatal hypoxia-ischemia (N = 28), and (5) an autopsy cohort of human infants (N = 78). Groups 1–3 are part of ongoing research investigations related to the prenatal consequences of maternal FLUAV or ZIKV infection using experimentally infected healthy pregnant dams (M. nemestrina) with either virus or saline (sham controls) at different stages of gestation with necropsy in the third trimester (Fig. S1, Table S1). Group 4 was added to review the fetal brain histopathology from neonatal macaques, which had been subjected to brief periods of perinatal hypoxia-ischemia caused by experimental clamping of their umbilical cords, as part of a previously published study (University of Washington IACUC permit #3328-05) (Table S2) [13]. Inclusion of Group 5 enabled a similar review of fetal brain histopathology from an autopsy cohort of 78 infants who died less than one month after birth (Table S3) in a deliberate effort to identify pathology like that observed in our viral-inoculation model.

Animal groups

An objective of our ongoing studies was to determine if the temporal course of maternal infection is associated with changes in fetal brain pathology. Therefore, we incorporated a time course into our ZIKV NHP studies to generate three study groups with different time intervals from maternal infection to delivery (Fig. S1): (1) Short (SHORT)-ZIKA (N = 5) inoculated in the third trimester and necropsied 2–7 days later; (2) Intermediate (INT)-ZIKA (N = 7) inoculated in the late second or early third trimester and necropsied 20–24 days later; and (3) Long (LONG)-ZIKA (N = 5) inoculated in first or second trimester and necropsied 43–97 days later. The LONG-ZIKA cohort was the basis for two prior publications significantly different in content from the current study [11, 12]. Within the FLUAV group, healthy pregnant dams were inoculated with FLUAV and Cesarean section and necropsy were performed 5 days later.

Viral inoculations

We subcutaneously inoculated 17 healthy pregnant pigtail macaques with ZIKV (5 × 107 plaque-forming units). For the SHORT-ZIKA and INT-ZIKA groups, a Brazilian ZIKV strain (Fortaleza, 2015) was used. For the LONG-ZIKA group, we inoculated a Cambodian ZIKV strain (F2213025, 2010) in 2 animals (ZIKA1, ZIKA2) and the Brazilian ZIKV strain (Fortaleza, 2015) in 3 animals (ZIKA3-ZIKA5). In the FLUAV group, we inoculated a pandemic strain of FLUAV H1N1 (A/California/07/2009) via four different routes (intratracheal, nasal, eyes, and oral) with a total inoculum of 7.4 × 106 plaque-forming units into 10 healthy pregnant pigtail macaques.

Animal care

All animals were monitored carefully for signs of illness and a battery of serological studies was performed on all dams prior to and during pregnancy to document infections with any endemic macaque pathogens; the details of the ZIKV, West Nile Virus, and dengue virus assays were described in our prior study [10]. Data was extracted from the University of Washington Primate Center Animal Research Management System database, which reflected historical testing of pathogens by the University of Washington Department of Laboratory Medicine or the WaNPRC Primate Diagnostic Services Laboratory. For many of the pathogens, testing was performed repeatedly before and during pregnancy and dams with any positive result were termed “positive” even if a later test was negative, which is the convention for test results for Monkey B Virus (Cercopithecine herpesvirus 1; CHV-1) and coccidioidomycosis (Valley Fever) (Table S4).

Viral RNA detection assay

We have previously published ZIKV real-time quantitative polymerase chain reaction (RT-qPCR) results from brain tissue samples of LONG-ZIKA fetuses using a ZIKV prME target which revealed positive results in ZIKA1 and ZIKA2 fetal brain [10]. In this project, we used a newly designed and optimized RT-qPCR targeting the ZIKV capsid with greater sensitivity and specificity. Viral RNA load was assessed in tissues from the dam, fetus, and placenta using a ZIKV or FLUAV H1N1 virus-specific RT-qPCR assay. Fetal and maternal organs were either immersed in RNA-later immediately upon harvest or flash frozen in Tissue-Tek Optimal Cutting Temperature (OCT) compound (Cat# 4583, Sakura Finetek USA Inc, Torrance, CA, USA) and were later weighed and homogenized in either QIAzol (Cat# 79306, Qiagen, Hilden, Germany) (for brain samples) or TRIzol reagent (Cat# 15596026, Thermo Fisher Scientific, Waltham, MA, USA) (for all other tissues). Tissues were homogenized using a Precellys Evolution bead-beater apparatus with the Cryolys Evolution cooling system. (Cat# K002198-PEVO0-A.0, Bertin Technologies, Montigny-le-Bretonneux, France). Maternal or fetal plasma was separated from aseptically collected dam or cord blood in heparin tubes through centrifugation at 1,500 x g for 10 min at 4 °C.

RNA was extracted from tissues using the RNeasy mini kit (Cat# 74106, Qiagen) and from serum or plasma using the QIAamp Viral RNA Mini Kit (Cat# 52906, Qiagen) according to manufacturer instructions. 800–1000 ng of RNA was used to synthesize cDNA using the iScript select cDNA synthesis kit (Cat# 1708897BUN, Bio-Rad Laboratories, Hercules, CA, USA) according to manufacturer’s protocols for gene-specific primers or random primers. Serum or plasma was treated with Bacteroides Heparinase I (Cat# P0735S, New England Biolabs, Ipswich, MA, USA) after cDNA synthesis.

Viral RNA was quantified using the TaqMan Fast Advanced Master Mix (Cat# 4444556, Thermo Fisher Scientific) and an Applied Biosystems QuantStudio 3 RT-PCR system (Cat# A28567, Thermo Fisher Scientific) with primers (Table S5) that correspond to residues conserved in both the FSS13025 and Brazil Fortaleza or H1N1 genome (GenBank numbers KU955593.10, KX811222.1, NC_026433.1). The thermal cycle program for ZIKV reactions consisted of one cycle of 120 s at 95 °C, then 45 cycles of 10 s at 95 °C, 15 s at 58 °C, and then 30 s at 60 °C to capture data. FLUAV reactions consisted of a single cycle at 50 °C for 120 s followed by 120 s at 95 °C, followed by 40 cycles of 1 s at 95 °C and 20 s at 60 °C where data was captured. Samples were considered positive for detection of viral RNA if amplification curves were less than or equal to a Ct (cycle threshold) value of 40. Copy number sensitivity, as determined using a standard curve from 7 tenfold dilutions of known quantities of ZIKV or H1N1 genome, was between 3 and 20 copies/qPCR reaction.

Plaque assay

The plaque assay was performed in two steps. The first step was to propagate live viruses from tissues or plasma by incubating plasma or tissues with a susceptible cell line. The ZIKV and FLUAV cohort samples were incubated with mosquito-derived C6/36 (Aedes albopictus clone, #CRL-1660, ATCC, Manassas, VA) or MDCK-SIAT1 (gift from Dr. Jesse Bloom) cells, respectively. The cell lines were routinely cultured in a solution of complete Dulbecco’s modified Eagle’s medium (cDMEM), comprised of DMEM (Cat# 10-013-CM, Corning, Corning, NY, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Cat# SH30396.03HI, Cytiva Life Sciences, Marlborough, MA, USA), at 28 °C. The C6/36 or MDCK-SIAT1 cells were plated in 6 well-plates (1,000,000 cells/well) with 10% FBS complete media at 28 °C overnight. When the cells were about 70% confluent, the ZIKV or FLUAV plasma samples or tissue homogenates were added to each well. The tissue samples were homogenized, spun at 1,500 rpm at 4 °C, and 500 µl of the supernatant was added to each well. For plasma samples, 100 µl of plasma was added to each well. For FLUAV plasma and tissues, the supernatants from each well were recovered after three days and spun at 1,500 rpm at 4 °C for 10 min. For plasma and tissues obtained from ZIKV cohorts, the media was changed after three days and plates continued incubating for 4 additional days. Supernatants were then recovered and spun at 1,500 rpm at 4 °C for 10 min. Then, the supernatants were aliquoted and stored in -80 °C for plaque assay.

In the second step, plaque assays were performed using either Vero cells (ZIKV cohort) or MDCK-SIAT1 cells (FLUAV cohort). Approximately 1 × 106 of Vero cells or MDCK-SIAT1 cells were seeded in six-well plates and incubated overnight in complete media (1x DMEM, 1X MEM Non-essential Amino Acid Solution, 10% FBS, 20 mM HEPES, 2 mM L-glutamine, and 100 U/ml penicillin and streptomycin). Serial dilutions were prepared in 1x DMEM and ~ 0.5 ml was added per well to the monolayer. The plates were then incubated for one hour (FLUAV cohort) or two hours (ZIKV cohort). After this absorption period, the viral inoculum was removed and the monolayer was washed with 1X PBS, and then 2 ml of media overlay (1.2% cellulose, 2X DMEM media) was added to each well. The plates were then incubated for 3 days (FLUAV cohort) or 5 days (ZIKV cohort) at 37 °C, until the plaques became visible, and fixed with 10% neutral buffered formalin for at least 2 h at room temperature. To visualize the plaques, the wells were stained with 0.5% crystal violet in 10% ethanol. The plaques in each well were counted. The results should be interpreted as semi-quantitative given the initial amplification step that was performed to enable detection of live virus at extremely low counts.

Necropsy and tissue sampling

At each macaque necropsy, the fetal cerebral hemispheres, cerebellum, and brainstem were divided in the mid-sagittal plane and one-half of the brain was fixed in 4% paraformaldehyde. The fixed halves were subsequently “bread loafed” to produce 10–11 coronal slabs (each ~ 4 mm thick) of the cerebrum and 3–4 transverse sections of brainstem / cerebellum, which were embedded in paraffin. Four micron-thick histological sections from these tissue samples were used for all histological and immunohistochemical studies. From the opposite half of the brain, separate RNA later and snap frozen samples were collected from cortical gray matter, cortical white matter, deep gray matter, right cerebellum, and right brainstem. In addition, ~1mm3 pieces of white matter lateral to the lateral geniculate nucleus and lateral to the frontal horn of the lateral ventricle were fixed in 4% glutaraldehyde in 0.1 M sodium cacodylate buffer. The dam’s brain was handled in a similar manner. Samples from all major organs of the fetus and dam were also snap frozen and processed for histology. The placental chorioamniotic membranes, fetal meninges and fetal blood were cultured to rule out bacterial contamination of the placenta or the fetus.

Histology, immunohistochemistry, and electron microscopy

Hematoxylin-and-eosin (H&E)-stained sections from every organ were evaluated. Except for visceral pathology observed in some of the FLUAV-inoculated dams, which will be reported separately, and the brain findings reported below, no significant consistent pathology was observed in the other dams, placenta, or any of the other fetal tissues. Coronal sections of the cerebral hemispheres were reviewed independently by two pathologists (R.K. and A.B.) with no significant discordant interpretation. Sections from many of the cerebral tissue samples were also pretreated with diastase and stained with periodic acid–Schiff–diastase stain (PASd) to highlight deposits of glycoproteins or glycolipids. Inclusion cells (ICs), as described in the results, were counted in each H&E-stained coronal section to establish the maximal number of ICs present in one coronal section from each brain.

Immunohistochemistry was performed with a Ventana Benchmark II automated immunostainer using the Optiview detection system (Ventana BenchMark Ultra; Ventana Medical Systems, Tuscon, AZ). The primary antibodies used in this study and relevant immunolabeling parameters are provided in Table S6. Briefly, we stained tissues with a battery of antibodies for IC characterization including: lysosomal-associated membrane protein 1 (LAMP1), lysosomal-associated membrane protein 2 (LAMP2), glial fibrillary acidic protein (GFAP), ionized calcium-binding adapter molecule 1 (IBA1), cathepsin S (CTSS), microtubule-associated protein 1 A/1B-light chain 3 (LC3), SRY-box 2(SOX2), SRY-box transcription factor 10 (SOX10), oligodendrocyte transcription factor 2 (OLIG2), myelin basic protein (MBP), marker of proliferation Kiel 67 (Ki67/MIB1), neurofilament, neuron-specific nuclear protein (NeuN), calretinin (CALB2), nucleoporin 62 (NUP62), caspase 3 (CASP3), and ZIKV non-structural protein 1 (NS1).

For LAMP1-, LAMP2-, and IBA-1-immunostained sections, optical densitometry was performed to quantify areas of immunoreactivity. Individual 200x fields of areas were photographed. Densitometry was performed by observers blinded to the pathogen-exposure history of each specimen. Digital Image Analysis (DIA) platform Visiopharm Integrator System (VIS; ver. 2023.01.1.13563; Visiopharm, Hørsholm, Denmark) was used to analyze the immunohistochemistry stains. Positive staining was detected by binary thresholding. The percent positive staining was calculated by comparing the area of the positive stain label to the whole tissue section area. Sections used for quantification of immunoreactive cells were counterstained with H&E. Other immunolabelled sections were counterstained with PASd to exclude or confirm colocalization of various antigens in the nuclei or cytoplasm of PASd-positive ICs. Distributions of ICs and areas of abundant or sparse LAMP1- and LAMP2-immunoreactive glial cells were mapped on drawings of coronal sections from the BrainInfo Macaque Atlas [14].

As controls for ZIKV immunohistochemistry, approximately 10 million virus-infected and uninfected Vero cells were collected from monolayers using trypsin enzymatic digestion. Cells and media were centrifuged at 800 rpm for 10–20 min and then washed with PBS. The cell pellets were then fixed with 10% neutral buffered formalin (NBF) for 24 h at room temperature. The cell pellets were then washed and resuspended with 70% ethanol at 4 °C for 24 h. Next, the cell pellets were spun for 10 min at 1,500 rpm and the ethanol was discarded and drained from the tube. The cell pellets were then added to the Epredia Cytoblock system (Cat# 7401150, Fisher Scientific) and prepared following manufacturer instructions. The lid of the Cytoblock cassette was closed and placed in 70% ethanol. The FFPE cell pellets were then embedded in paraffin and sectioned for staining and viewing.

Glutaraldehyde-fixed samples of deep white matter were washed 5 times for 5 min each time in buffer at room temperature and post fixed in buffered 2% osmium tetroxide, on ice, for 1 h. After 5 washes in distilled water and en bloc staining in 1% uranyl acetate overnight at 4 °C, the tissue was washed 5 × 5 min in water, dehydrated in ice cold 30%, 50%, 70%, and 95% ethanol, and then allowed to come to room temperature. This was followed by two changes of 100% ethanol, 2 changes of propylene oxide, and infiltration by a 1:1 mixture of propylene oxide: Epon Araldite resin for 2 h. Next, fresh Epon Araldite was exchanged with the mixture twice (2 h per change) and the tissues were embedded in Epon Araldite at 60 °C overnight. The tissues were cut to yield 80 nm sections that were stained for 2 min in Reynold’s lead citrate and imaged with a transmission electron microscope (JEOL USA, Peabody, MA, USA) at 80 KV.

Autofluorescence

Autofluorescence was assessed in 5-µm-thick paraffin sections from IC-rich brain tissue along with sections of lipofuscin-containing macaque cerebellum as a positive control. Sections were deparaffinized and incubated for 10 min in a dilute solution of 4′,6-diamidino-2-phenylinverdole (DAPI; 5 ul in 50 ml water), rinsed, and coverslipped in water. Photomicrographs of autofluorescence (green– excitation 467–498 nm and red – excitation 542–582 nm) and DAPI (excitation 352–402 nm) were taken and then the coverslip was removed and PASd staining was performed. The same microscopic fields were identified and rephotographed under brightfield optics.

Review of brain slides from human cases and other models of brain injury

A review was conducted of H&E-stained coronal sections from the brains of neonatal macaques, which had been subjected to brief periods of perinatal hypoxia-ischemia caused by experimental clamping of their umbilical cords, as part of a previously published study (University of Washington IACUC permit #3328-05) [13]. Details concerning the experimental manipulations and some of the findings in these animals are provided in Table S2. A similar review of H&E-stained brain slides was performed in an autopsy cohort of 78 infants, who died less than one month after birth (Table S3). Paraffin sections from the same tissue blocks were prepared from a subset of these patients and either stained with PASd or immunohistochemically, as described above. These cohorts were included in a deliberate effort to identify ICs like those observed in our pregnant NHP viral inoculation models.

Statistics

Fetal inclusion cells were analyzed as a continuous variable and binary variable using different thresholds. When fetal inclusion cell positivity was considered a binary variable, we used the Fisher exact test or the Kruskal-Wallis test. When the analysis was performed as a continuous variable, we used the Spearman rank-order correlation test. A p-value less than 0.05 was considered significant.